Click here or press enter for the accessibility optimised version

Ghost in the
Machine:
An IT pro’s definitive
guide to generative
AI
 What Is AI-
Why Assisted
Generative AI: Microsoft Coding, and
A Copilot Should You
Cybercriminal’s Technology Use It?
What Is New Best Will Change
Can ChatGPT ChatGPT? Friend How We Work Read More
Write How It Works
PowerShell and Best Uses Read More Read More
Code? for Chatbots

Read More Read More

Click here or press enter for the accessibility optimised version

Can ChatGPT Write PowerShell Code?

We put ChatGPT’s PowerShell scripting abilities to the test. See how the AI chatbot performed.
 Test #1: A Basic PowerShell Script

As amazing as ChatGPT is, I

recently heard several developers
express concern that the platform
might eventually put them out of a
job because of its ability to write
code. Conversely, other developers
have said that ChatGPT is of no
concern because ChatGPT writes Figure 1. ChatGPT gave me instructions for writing a Hello World script.
code poorly.

The simple fact that developers are I began my tests. I asked ChatGPT to help me to The experiment shown in Figure 1 convinced me
talking about ChatGPT’s code- write a simple “Hello World” script using that ChatGPT can indeed write PowerShell
writing capability got me thinking PowerShell. As you can see in Figure 1, ChatGPT scripts. Even so, explaining how to use
about PowerShell. Specifically, provided instructions for creating a Hello World PowerShell for performing a simple task isn’t the
could ChatGPT write PowerShell script, right down to explaining how to open same as generating a usable script. As such, I
code? If so, are its scripts well PowerShell. decided to do another experiment.
written?
Test #2: More Complicated PowerShell Code
This experiment involved writing a PowerShell For my test, I wanted to create a more I decided to ask ChatGPT to create a simple
script myself and then asking ChatGPT to write complicated script than the Hello World script script to count from 1 to 100, wait 250
the same script. I would compare the scripts. I yet not super-complex. After all, my goal was to milliseconds between each number, and display
didn’t worry about trying to see who could write compare AI-generated code against my code, a progress bar.
the script fastest. Even though I work with and that can be accomplished without writing a
PowerShell every day, ChatGPT would almost long, arduous script.
certainly write the code more quickly than I could.
Here is the PowerShell
code that I wrote:
For ($Count = 1; $Count -le 100; $Count++
)

Write-Progress -Activity “Counting”

-Status “$Count% Complete:”
-PercentComplete $Count

Start-Sleep -Milliseconds 250

Figure 2. Here is what it looks like when I run my script.

Click here or press enter for the accessibility optimised version

What Is ChatGPT? How It Works and Best Uses for Chatbots

ChatGPT stands out from conventional chatbots for its general-purpose nature. Learn about ChatGPT’s
benefits and how to create a custom chatbot
ChatGPT is best described as an AI-driven chatbot that stands
out from conventional chatbots for its general-purpose nature.
While most chatbots are designed for specific purposes,
ChatGPT is versatile and can handle a wide variety of tasks
effectively.
How Does ChatGPT Work? Access to a wide range of datasets
At its core, ChatGPT combines a large language model and an AI engine that has undergone training ChatGPT has been extensively trained on a
on a massive dataset. The dataset includes information from countless websites, books, articles, massive dataset, making it knowledgeable on
and other sources. various topics of interest to website visitors.
Incorporating ChatGPT on the backend allows
During the training process, ChatGPT’s underlying model was constructed in such a way that it can the chatbot that you build to tap into the
receive a question as input. It then uses the information it gained from the training data to generate knowledge base acquired during its training.
a relevant response.

Benefits of Using ChatGPT Enhance the accuracy of results

While ChatGPT is designed for general- Traditional chatbots are limited to
purpose use, it is possible to build a custom preprogrammed responses. To be useful,
chatbot that uses ChatGPT as the backend. programmers must anticipate and provide
Incorporating a ChatGPT-based chatbot on a answers to the questions that site visitors are
website or web application offers a multitude likely to ask. Over time, these answers may
of advantages. become outdated.
Ability to handle complex questions
Increase customer engagement By using ChatGPT, however, website owners can ChatGPT excels at handling complex queries.
Integrating ChatGPT on a website or web improve the chances of visitors receiving up-to- Unlike traditional chatbots that rely on users
application can help to enhance customer date and precise information in response to their phrasing questions in just the right way, ChatGPT
engagement. With widespread adoption, queries. can process and respond to intricate questions –
ChatGPT provides a familiar user experience. Its even if they contain misspellings or poor
conversational nature encourages site visitors to phrasing. This increases the odds of website
stay longer and actively engage. visitors receiving satisfactory answers to their
questions.
 Step 1: Choose your chatbot language Step 4: Configure the bot settings
Select the programming language that you want Customize your chatbot’s settings to fine-tune its
to use. Calls to ChatGPT will require using responses, optimize resource consumption, and
OpenAI’s API, meaning that you can use any improve response time.
programming language you like so long as it can
generate HTTP requests and parse JSON Step 5: Create an action plan for users
responses. Develop a user action plan that guides users on
how to interact with your chatbot effectively. This
Step 2: Choose a voice engine might involve explaining the bot’s purpose and
Although not an absolute requirement, consider basic functionality. You might highlight its
integrating a voice engine to enable speech primary functions or provide guidelines for
capabilities for your chatbot. When doing so, engaging in conversations with the bot.
evaluate the quality of speech output,
naturalness of the voice, multilingual support, Step 6: Build an audience offering

How To Set and ease of integration with your chatbot. Tailor your chatbot’s behavior and characteristics
to target a specific market segment. Align the

Up and Use a Step 3: Pick a personality model

Depending on your chatbot’s purpose, select a
chatbot’s traits with the preferences and needs
of the audience you aim to target.

Chatbot With personality model to determine its overall

demeanor. For example, you can decide whether Step 7: Promote your bot

ChatGPT your chatbot should have friendly, professional,

humorous, authoritarian, or other specific
Remember, all your hard work developing your
chatbot will have been for nothing if no one uses
personality traits. the bot. As such, make efforts to generate
To set up and use a ChatGPT-based awareness and encourage users to engage with
chatbot, follow these seven steps the chatbot you created.
Frequently What are the best uses for ChatGPT?
The potential uses for ChatGPT are limited only by your imagination, as it offers versatile capabilities.

Asked Common applications for ChatGPT include content creation, writing programs, trip planning, and
general research. Ultimately, ChatGPT can be used for a wide range of purposes.

Questions
What type of data does ChatGPT use?
ChatGPT has been trained on a vast amount of data, including web content, books, articles, and
scientific papers.

Does ChatGPT use artificial intelligence?

Yes, ChatGPT is based on artificial intelligence. ChatGPT falls under the category of generative AI,
where an AI engine creates content based on the training.
 How is ChatGPT different from other chatbots?

Frequently The main thing that separates ChatGPT from other chatbots is that it is general-purpose in nature.

Asked Before ChatGPT’s release, most chatbots had specific uses. For example, an organization might use a
chatbot on its website as a tool for answering frequently asked customer questions. ChatGPT’s training

Questions on a vast amount of data lets it provide answers to questions on various topics.

Are there any risks associated with using ChatGPT?

The main risk of using ChatGPT is the possibility of receiving incorrect answers. Even though ChatGPT
is amazingly capable, it can and sometimes does generate inaccurate responses. Similarly, if you ask
ChatGPT to create a software program for you, you will likely be required to do debugging before the
program becomes fully functional. It’s also worth considering that ChatGPT’s training was completed in
2021, so it has limited knowledge of events occurring after that time.

What are some of the most popular ChatGPT applications?ChatGPT can be extended through plug-
ins, and some of the most popular plug-ins include Expedia, Instacart, and OpenTable.
Click here or press enter for the accessibility optimised version

Generative AI: A Cybercriminal’s New Best Friend

With the rising popularity of generative AI, such as ChatGPT, cybercriminals are likely to exploit its capabilities for
convincing phishing scams and in-depth vulnerability research.
 A New Tool for Phishing Scams
First, I believe cybercriminals will use generative received a phishing email that claimed to come
AI to create convincing social engineering from my bank, but whoever wrote the email
scams, the most likely being phishing email misspelled the bank’s name and my name. As if
In recent months, ChatGPT and other messages. that weren’t bad enough, they even misspelled
generative AI have grown massively the word “account.”
popular. However, history has conclusively It’s safe to say that we have all seen obvious
shown that anytime a particular phishing emails, some that made us laugh. I’m With that in mind, imagine phishing scammers
technology goes mainstream, talking about messages that claim to come from who now have ChatGPT at their disposal. They
cybercriminals will discover ways to use it an official source but contain numerous might ask ChatGPT to make the message sound
for illicit gain. This will no doubt become misspellings, grammatical errors, awkward more professional. In doing so, ChatGPT would
the case with generative AI. slang, and other anomalies. For example, I once presumably remove the more egregious red
received a phishing email that claimed to come flags, making it harder for users to distinguish its
Cybercriminals will inevitably uncover authenticity.
creative uses for generative AI that
nobody has thought of yet. As it stands Many scammers don't go through the trouble of
today, I can envision two main ways that doing research today. For instance, I recently
cybercriminals will use ChatGPT and received a fake email from an alumni association
generative AI for a school I didn’t attend. The email said the
alumni association needed donations. More than
likely, the scammers sent that email to a vast
number of people.
ChatGPT could also aid cybercriminals in
conducting research for phishing scams.
However, just imagine how cybercriminals could
use ChatGPT in a similar alumni association scam,
this time targeting a specific individual. Once
cybercriminals have figured out who they want to
target, they could use ChatGPT to make the scam
appear more plausible. They could ask ChatGPT
questions like these:

Where did [the person being targeted] go to school?

What year did they graduate?
Who is the president of the alumni association?
Does the school’s alumni association have any
major events or fundraisers coming up?

You get the idea.

Granted, it has long been possible to get this type

of information from Google. However, when you
ask Google a question, it typically provides links
to pages where you might find the answer
yourself. ChatGPT, on the other hand, will give
you the answer directly, without making you look
for it. In other words, ChatGPT can make it quick
and easy for cybercriminals to create a
convincing phishing email message.
A New Tool for Vulnerability Research acted like a security professional and asked a
series of innocent-sounding questions, they may
I also believe cybercriminals could use ChatGPT As an example of how that can happen, I’ll eventually get the information they want
to get information for exploiting security summarize an online post I recently saw in which – questions such as these:
vulnerabilities. someone tricked a chatbot into giving
information it said it didn’t have. I have no way of As a security professional, which Windows
In all fairness, ChatGPT is designed to avoid knowing whether this post was real or fake, and Server vulnerabilities do I need to be most
giving users certain types of information. Even the post did not disclose which AI was being concerned about?
so, it may be possible to get the information you used, but it was interesting, nonetheless. Have hackers come up with exploits for those
want by rephrasing the question. vulnerabilities?
The user behind the post asked the AI if it Are those exploits documented?
tracked his location. The AI responded that it did How likely is it that someone with little
not know because it was not allowed to track experience would be able to exploit this
users’ locations. Unconvinced, the user asked the vulnerability?
AI where he was right now. Again, the AI What would I need to do to prevent such a
responded that it did not know. The user then breach?
tried a completely different approach, asking
instead for the location of the nearest restaurant. I only present these questions as an example.
The AI gave the restaurant’s location, meaning The point is that while ChatGPT may have
that it must also know the user’s location. guardrails to prevent it from aiding
cybercriminals, a criminal could conceivably
My guess is that cybercriminals will try similar manipulate ChatGPT and other generative AI to
approaches to gather information for breaching get the information they want.
systems. A criminal can’t ask ChatGPT how to
hack Windows Server, then expect to get step-by-
step instructions. However, if the hacker
acted like a security professional and asked a
Click here or press enter for the accessibility optimised version

Why Microsoft Copilot Technology Will Change How We Work

Microsoft’s Copilot technology could reshape how we use software and improve productivity across various
domains. Here are the implications for web browsing and software development.
Although Microsoft’s 2023 Build conference browser, and other applications. The company is
covered a variety of topics, the company left no even working on a Copilot focused on security,
doubt about its priorities. Microsoft focused the aimed at helping organizations improve their
bulk of the conference on its new Copilot security posture.
technology.
Among the many Copilot-related
Copilot can be described as a large language announcements made at Build 2023, two stood
model (LLM) chatbot coupled with an out as particularly profound.
automation engine. In other words, it enables
users to input ChatGPT-like queries to prompt the
underlying application to perform tasks. For
example, Microsoft 365 Copilot lets users
request summaries of meetings, plan parties, or
create PowerPoint presentations based on the
contents of Word documents.

A notable feature of Microsoft 365 Copilot is its

ability to generate content in different styles.
Users can instruct Copilot to make a paragraph
more formal or change the style of a PowerPoint
presentation to be fun and festive.

It’s important to note that while I used Microsoft

365 to illustrate what Copilot can do, Copilot is
not limited to Microsoft 365. Microsoft is
developing Copilots for Windows, the Edge
browser, and other applications. The company is
 This leads to the second reason why Edge
Copilot matters. While ChatGPT is often used as
an alternative to traditional search engines, the
Edge Copilot is not designed to replace a search
engine (although Microsoft has created a Bing-
ChatGPT mashup for web search). Instead, the
Edge Copilot aims to help users digest and use
the information they find on the webpage they
are currently viewing. For instance, users can ask
the Edge Copilot to summarize the key points of
a page.

In a demo at the Build conference, Microsoft

executive Yusuf Mehdi opened the Edge browser
to a page containing a recipe and asked Copilot
to list the recipe’s ingredients and add them to
Edge Copilot Will Transform Web Browsing Instacart. The demo illustrated Edge Copilot’s
ability to analyze webpage contents (in this case,
One of the significant announcements was the information about events occurring after 2021. identifying recipe ingredients) and work with
development of a Copilot for Microsoft Edge. As such, ChatGPT usually isn’t the best tool for extensions (adding the ingredients to Instacart).
There are two compelling reasons why a browser gathering information on the latest technologies My guess is that Edge Copilot will continue to
Copilot holds importance. or events. However, the Edge Copilot will have enhance the usefulness of the browser in the
access to the same up-to-date data as the Edge future.
Firstly, the Edge Copilot will bring ChatGPT-like browser, enabling it to use information from
functionality to current information. A major webpages regardless of when the page was
drawback of using ChatGPT today is its limited created, be it five years ago or 20 minutes ago.
information about events occurring after 2021.
Copilot SDK Will Transform Software Development Ultimately, the specific uses and the companies
that will develop Copilots remain to be seen.
The other Copilot-related announcement that Now, imagine if the CAD software had contained Nevertheless, I believe that Copilot has the
caught my attention was Microsoft’s plan to a Copilot. I could have used the Copilot to assist potential to revolutionize the way we use software.
release the Copilot SDK to the public, allowing me with the more technically complex aspects of
third-party developers to build their own Copilots. my project. For example, instead of manually
As more developers embrace this technology, I adjusting the height of my living room ceiling, I
anticipate that working with Copilot will become could have simply asked Copilot to raise the
increasingly accessible. ceiling by four feet.

To illustrate the potential impact, let me share an Copilot has enormous potential to simplify how
example from my personal experience. Last year, we interact with complex applications. However,
I used a popular CAD program to create a 3D there is another potential use case that I have yet
model of my home. I wanted to undertake a to hear anyone discuss. As previously mentioned,
home improvement project and ensure its Copilot combines an LLM chatbot with an
feasibility before purchasing materials. The 3D automation engine. Consider the possibilities if
model let me explore various options before multiple developers collaborated to create cross-
committing to the project. product Copilots. Suppose a video editing
software company partnered with the
As useful as my CAD model was, creating it was aforementioned CAD software company. With a
no small task. I encountered numerous instances complete CAD drawing, Copilot could be used to
where I made mistakes and had to painstakingly instruct the software to generate a 3D
backtrack to make corrections. There were also walkthrough, export the walkthrough to the video
times when I struggled to find solutions and had to editing software, add background music and color
watch many YouTube tutorials. grading, and export the final video as a 4K MPG
file.
Click here or press enter for the accessibility optimised version

What Is AI-Assisted Coding, and Should You Use It?

Should developers use AI to generate code? Here are the pros and cons of AI-assisted coding.
 What Is AI-Assisted Coding? a simple form of automated code generation.

AI-assisted coding is the use of machine learning What's different about AI-assisted code
models to generate code automatically. Most generation, however, is that, like low-code and
existing AI-assisted coding tools let developers no-code tools, it doesn't rely on prebuilt code
describe in natural language what they want their modules to allow developers to construct
code to do. Then, the tools automatically applications without writing much (or any) code
generate the code to do it, complete with custom themselves. Instead, AI-assisted coding tools
For years, the idea that you could variable names and the invocation of external generate custom code entirely from scratch.
automatically generate custom modules or services, where necessary.
source code seemed like fiction. This means that, at least in theory, there's no
But it has become reality thanks to In the background, the tools are powered by limitation on what AI-assisted code generation
new services like GitHub Copilot algorithms that were trained using vast amounts tools can do. As long as the algorithms are good
and Tabnine, which provide AI- of existing source code — which usually comes enough and the data they are trained on is
assisted code generation. in the form of publicly available source code comprehensive enough, AI-assisted coding tools
produced by open source projects. Based on could in principle write highly complex
Solutions like these have shifted these examples, the algorithms generate custom applications based only on natural-language
the question from "how can we use code on demand. descriptions of how those applications should
AI to generate code?" to "should we work.
use AI to generate code?" How New Is AI-Assisted Code
Generation? Although developers have long talked — usually
There's no clear answer, but there is The idea of automatically generating source half-jokingly — about the idea of using AI to write
plenty to say about the pros and code for applications isn't especially new. Low- code, AI-assisted coding solutions have really
cons of AI-assisted coding in code and no-code development tools have been only matured over the past couple of years. They
today's development environment. around for decades. And IDEs have long offered represent perhaps the latest major innovation in
autocomplete features, which you could consider the realm of software development
a simple form of automated code generation.
Where to Find AI-Assisted
Coding Tools
A variety of vendors now sell AI-assisted
coding tools or services. Probably the
most famous are GitHub Copilot (which is
based on OpenAI Codex) and Microsoft
Intellicode. Smaller companies, including
Kite and Tabnine, also offer production-
ready tools for AI-assisted code
generation.

Open source tools exist in this domain,

too, although — like many open source
solutions in newly established domains —
they tend to be rougher around the edges
than the commercial alternatives. To date,
the main open source AI-assisted coding
tool is PolyCoder.
The Pros and Cons of AI-Powered Coding
AI-assisted code generation offers one main benefit: It saves developers time by allowing them to write code without actually having to write code.

You could also argue that AI-powered coding tools are helpful in situations where developers want to implement certain code but don't know how. I'm not
convinced that that is the case currently, however, because AI-assisted coding tools aren't sophisticated enough — at least now — to be able to generate truly
complex code. Plus, to get good results when using these tools, developers often have to be relatively specific about how they want their application to do
something — so developers more or less need to know how they would write the code, even if they use AI-assisted tools to write it for them.

On the other hand, there are clear drawbacks to AI-assisted coding:

Low accuracy: Advanced AI-assisted Extensive code review: Developers have Legal and ethical issues: One of the
coding solutions, like OpenAI Codex, to accept (or reject) automatically stickier issues involving AI-assisted coding
generate accurate code only 37% of the generated code as they work. For some tools is that, because the tools are trained
time, according to their developers. That coders, the need to review code constantly using open source code, they essentially
obviously falls far short of the could be distracting. They may be able to reproduce code that was written by
performance necessary for programmers work more efficiently if they just write out another project. This raises legal and
to let AI fully take the wheel (which, to be all of their code manually. ethical questions, which have not yet been
fair, is not what these solutions currently sorted out. If your AI-assisted coding tool
promise — there's a reason why Copilot is produces code that closely resembles
not named Autopilot). code in another GitHub repo, does it make
you a plagiarizer or a license violator? We
don't yet know, but we probably will gain
perspective as people begin suing each
other over these issues.
These challenges mean that AI-assisted coding scrutiny, and who therefore are less likely to get know how good AI-assisted coding tools end up
tools may be handy for developers who want a caught up in accusations that they've stolen being, and whether they expose users to
fast way to generate relatively simple code, and other projects' code through their AI-powered complicated legal or ethical challenges.
who are willing to review — and, if necessary, code generators.
tweak — it to ensure it's accurate. I also suspect
that we'll see AI-assisted coding used more often But time will tell. For now, AI-assisted coding
by programmers working on proprietary projects represents a frontier that is still being conquered.
whose source code won't be exposed to public It will probably take several more years before we
scrutiny, and who therefore are less likely to get know how good AI-assisted coding tools end up
Click here or press enter for the accessibility optimised version

Thank you!

ITPro Today is your

home for news and
resources on IT
operations and
management.
https://www.itprotoday.com/

Cookies Terms Privacy POWERED BY