Curious about how ChatGPT can help you at work? Be careful what company information might be given away while experimenting, warns UNSW Business School professor, Rob Nicholls.

OPINION by Rob Nicholls >>

WHETHER IT’S MILD CURIOSITY or the business imperative of starting to use artificial intelligence (AI), as part of a unique selling proposition, many people now use ChatGPT, either at work or at home.   

When it launched, the user-friendly AI large language model went from zero to 100 million users in only 60 days. Since then, Microsoft has invested $US10 billion dollars in start-up Open AI and embedded GPT4 into Bing. Dedicated Bing apps with GPT4 are now available on the App Store and the Play Store.  

Part of the user experience of OpenAI’s ChatGPT is that AI tools can generate really useful text from a specific prompt, making for possible opportunities to save time in the day-to-day at work with machine learning – for example, when writing emails.  

You can even use the job description provided in an advertisement to get ChatGPT to write the ‘perfect’ cover letter. Or the staff member from People and Culture can generate the “perfect” job advertisement for LinkedIn from the scratchy brief provided by the business unit.   

Trying out ways ChatGPT can be used comes with some risks at work

As a starting point, let’s look at the example of the job advert.  

If the role is one which is commonly advertised, then little is lost by sharing the form of the job description with a couple of hundred million other uses of ChatGPT. 

But if the job description includes information which could be used by a competitor to identify your business, then the risks are significantly higher (especially if recruitment is an important part of insider business strategy in your workplace).   

With companies like Samsung having recently been stung by staff members inadvertently giving away material via ChatGPT, it is important to consider the risks carefully before using it at work.  

And just like that – the code was gone  

This issue is particularly challenging if staff members have started using ChatGPT as part of the code development process.   

It’s really appealing to do so. One of the great aspects of ChatGPT is its automation use case in reducing coding time in software development projects for programmers.

This can be done at a design level, “How do I sort job application letters using Natural Language Processing in Python?” or at the code level, “How do I use Gaussian Naive Bayes in scikit-learn?”.   

(One really useful feature is the ability to paste code in as a prompt to ask ChatGPT to improve it. This is as simple as “what is wrong with the following code?”. ChatGPT can even recognise the coding language that you’re using!)  

The problem is that ChatGPT can then include the material that you have used as a prompt to improve its answers in the future, ‘training’ the algorithms. This is precisely what happened to Samsung developers who used ChatGPT to both improve code and keep meeting notes.  

Material that would have been regarded as some of the most sensitive by Samsung was available to developers outside of Samsung, simply because Samsung engineers used ChatGPT to decrease their development time.   

Will GPT4 make using ChatGPT at work riskier?  

The latest in the GPT series after GPT3, GPT4 has an incredibly accurate voice-to-text feature. But the risk of the text becoming part of the training set to improve the generative AI is high.  

It’s a simple way to both transcribe work meetings and, even better, to generate the minutes before the end of the meeting. However, the transcription will also be part of the GPT4 ecosystem before the end of that meeting. There have been news headlines about Italy’s decision to ‘ban’ ChatGPT over privacy concerns.   

Essentially, the argument made by Italian authorities is that the data collected by ChatGPT was in breach of the European General Data Protection Regulation. However, and consistent with other European countries, it seems likely that Italy will walk back from this approach by the end of April. The only change required will be to have an age verification (over 18) check on users.  

Generative AI uses billions of data points in order to be able to create text on a predictive basis. It improves in response to user feedback. The challenge faced by businesses that employ curious people is that this feedback may include company confidential material.   

The solution is simple in theory, but much harder in practice. If material would not normally be disclosed outside of the business, it should not be used as a prompt for ChatGPT or for Bing.   

The practical difficulty with this rule is that search engines, including Google with its generative AI called Bard, are an essential business tool. The issue may be to decide whether search engines are there to provide information, or to provide answers.  

So, I should avoid using ChatGPT at work?  

Not sure what you should (and shouldn’t) share with our friend, ChatGPT? Try this simple test:  

Is the output of the ChatGPT session a document that would normally be regarded as confidential by your business? Then it should not be shared on ChatGPT.  

If you did write your cover letter or resumes using ChatGPT, the AI system used to filter applicants could also run your cover letter through GPTZero. This online tool from Open AI can detect whether text was written by a generative AI by examining that text’s ‘perplexity’ (a measurement of the randomness of the text) and ‘burstiness’ (a measurement of the variation in perplexity).  

But of course, the improvement in ChatGPT’s text output is challenging these tools. So, who is to say how this will change into the future with developing AI technology? 

About the author

Rob Nicholls is an associate professor at the School of Management and Governance, University of New South Wales (UNSW) Business School.

ends

By Leon Gettler, Talking Business >>

CYBER SECURITY has become challenging now, with remote work and people working off any device.

Citrix Australia and New Zealand managing director, Martin Creighan said the problem now is that people’s personal, work and tech worlds have become more complex.

“We’re using mobile phones that have more computing power than the first space shuttle,” Mr Creighan told Talking Business.

“We’re using computers on a day-to-day basis. We’re connecting to different systems, to various organisations. We are working now from multiple places, so people are enjoying hybrid or remote work. They’re working from home or they’re working from the café down the road or they’re working from the office.

“All of that makes what we’re doing, and the job for our cyber security teams, even more complex.” 

Mr Creighan said remote work, with workers using any device, made things more complex and difficult for IT managers and chief information security officers.

“We’ve seen a proliferation over the last couple of years, as COVID hit, and everyone going remote and bring your own device to work,” Mr Creighan said. “And with multiple devices out there – devices that may not have been tested on the corporate network, or coming in and accessing the network from a wi-fi connection, or a connection that may not be secured at a corporate check.

“It just increases the complexity that IT managers have to deal with and that’s a big struggle for them.”

SURVEY BRINGS PROBLEMS TO LIGHT

As part of Citrix Australia and New Zealand’s efforts to come to terms with the new work challenges, the company conducted a survey of 250 IT leaders in Australia.

The leaders were asked whether their expenditure on cyber security would decrease with soaring inflation and the rising cost of living.

All of them said there would be no decrease. Organisations would keep spending on cyber security. They were saying that data protection, IT security and innovation would remain key areas of spending.

Most said they would spend at least as much, if not more, on these areas.

“And 63 percent of Australian IT leaders overall expect an increase in spending on data protection,” Mr Creighan said.

MEETING A CHALLENGING WORLD

The increased spending indicated in the business leaders survey was in response to a world that was becoming more difficult to manage.

“The challenge is how to do that in a world where it’s becoming more complex, because we have more people accessing our data and other sources of data from multiple devices in multiple places at multiple times,” Mr Creighan said.

He said a lot of cyber security incidents were the result of people not paying attention to what they were doing. So they click a link they think will take them to an important application or a fake banking application link.

As a result, he said, 63 percent of security breaches were caused by human error.

The other big challenge for companies and cyber security professionals was that the cyber ‘crooks’ were now using advanced technology, artificial intelligence, robotics and machine learning to hack into systems.

“It’s this cat and mouse game of us trying to leap frog one another,” Mr Creighan said.

“As we get more tools and processes and systems in place, they’re using the tech to do bad things and we’re using the tech to try and protect ourselves.”

www.citrix.com

www.leongettler.com

Hear the complete interview and catch up with other topical business news on Leon Gettler’s Talking Business podcast, released every Friday at www.acast.com/talkingbusiness.

https://play.acast.com/s/talkingbusiness/talking-business-40-interview-with-martin-creighan-from-citr

ends

CYBER PROTECTION group Nozomi Networks is marshalling Australian online security talent to join a first-of-a-kind series of simulated threat hunts, to protect key national infrastructure from attack, in a cyber war game scenario on December 1.

Nozomi Networks Inc. has created the event -- named Operation Cyber Crisis -- to bring together Australia’s best cyber talent to strategise in simulations how to stop cybercriminals aiming to take down key critical infrastructure.

Operation Cyber Crisis will play out two increasingly difficult cyber threat scenarios targeted against industrial systems and critical infrastructure. The challenges will incorporate elements of cyber reconnaissance, operations security, and malware analysis, putting the teams in tough scenarios to truly assess their decision-making abilities under stress. Participants will be sent a challenge coin after the event. 

The event will be run and managed by retired United States Marine Corps (USMC) Colonel Bill Hagestad II, a China cyber espionage expert who has worked extensively with the FBI and the SANS Institute.

“Operation Cyber Crisis is more than just fun, it’s about testing your cyber skills and building up your preparedness in the face of an attack,” Colonel Hagestad said.

“The crisis control team has formulated the game based on real-world threats and are on hand to analyse and evaluate each move the teams make. It’s a game of strategy and skill, but most importantly it shows the participants how their cyber skills will hold up when confronted with a serious threat.”

The games will be an opportunity for Australia’s top cyber minds to flex their problem-solving muscles and demonstrate and assess the best ways to counter incoming attacks.

Practising for real-world events builds up specialists’ cyber resilience, helping them be better prepared for the serious and rapidly evolving cyber landscape, according to Nozomi Networks regional manager for Australia and New Zealand, Ameen Al-Majzoub.

“As Australia looks to fill up to 30,000 cyber roles in the next few years, quality of training at all levels becomes so important," Mr Al-Majzoub said.

"This is a series of lifelike scenarios combining multiple attack methodologies together and will be an invaluable experience for any cyber pros looking to gain an understanding of what can happen in the trenches.”

Other members of the crisis control team will include former Queensland Police operational commander of the Fraud and Cybercrime division, Brian Hay; former US Department of Justice Federal Prosecutor, Jonathan Rusch; and Nozomi Networks’ director of cybersecurity strategy, Chris Grove.

The virtual event will be held on Thursday, December 1, AEDT from 10am to 12pm.

Interested participants and spectators can register here.

 www.nozominetworks.com

ends

THE RECENT MyDeal incident is the third major cybersecurity  ‘fail’ affecting Australians. University of NSW (UNSW) digital experts are calling for Australian business to restore public confidence in Australia’s cyber capabilities – fast and effectively.

In the past three weeks, three cyber attacks resulted in identity leaks that have sent ripples around the nation and raised the concerns of UNSW Institute for Cybersecurity (IFCYBER) chief scientist and professor, Sanjay Jha.

“Has the triple-A of cyber security (authentication, authorisation and accounting) failed?” he asked.

In the recent MyDeal.com.au incident, early reports suggested that a compromised credential, most likely related to elevated users, was used to access a database storing user information. The attack led to the scrambling of 2.2 million customers’ data, which included email addresses, full names, phone numbers, delivery addresses and some customers’ birthdates. 

In an interview with Channel 9 News, Professor Jha said, “… the breach raises serious concerns for the end user's confidence in using online services and poses a serious challenge for the industry”.

“Compromised credentials should not provide easy access to malicious actors when multi-factor authentication (MFA) is in place,” he said.

“You would expect more stringent authorisation and access control and network partitioning to protect these critical assets.”

SAFELY NAVIGATING THE DIGITAL WORLD

Following basic practices on Cyber Security Guidelines listed by the Australian Cyber Security Centre can help a business to protect its systems from cyber threats.

Arash Shaghaghi, a senior lecturer in cybersecurity from the UNSW School of Computer Science and Engineering and UNSW Institute for Cybersecurity, said it was evident from the recent attacks that some industries were not adopting the latest research on cybersecurity fast enough.

“Users’ data is collected by various services without any control from the end user, and often the data collection to join these services is excessive – leaving end users vulnerable and with limited options when a serious breach occurs,” Dr Shaghaghi said.

“Other parts of the world are investing heavily in technological measures such as self-sovereign identity (SSI), where users would have better control over who had access to what part of user identity and other information. 

“SSI gives individuals control over the information they use to prove who they are to websites, services and applications across the web,” Dr Shaghaghi said.

“We need to enhance investment in practical research and think of measures that facilitate the adoption of the latest technologies to reinforce our resiliency against the growing number of attacks targeting Australia.”

THE OPTUS CASE

Last month, 10 million Optus customers experienced a similar fate when cyber criminals hacked into the system, stealing personal details such as passport ID and driver licence numbers. It was reported in the media that the breach suggested an open port without authentication, for testing purposes, was left in the production version.

Unfortunately, these problems are well-known to the industry, Prof. Jha said.

“About five years ago, we were white boxing early versions of Philip Hue bulbs and Phillip Hue Bridge. We found that you could control these devices through such open ports,” he said.

“This was reported to the company who fixed the problem in their future version.  A simple penetration testing of servers before deployment could have potentially made such attacks difficult, if not impossible.” 

Prof. Jha is also concerned about the state of risk assessment and authorisation processes across the industry and said further stringent penalties for negligence would go a long way.

Often these processes were a box-ticking exercise in an Excel spreadsheet, and many conducting these tasks do not have adequate background in cybersecurity, he said.

“This re-emphasises the need for quality education along with more research in quality tools to improve these processes,” Prof. Jha said.

“My team is working on such tools for a Distributed Energy Resource Management Security project at UNSW funded by Cyber Security Cooperative Research Centre.

“Cyber security is a cat-and-mouse game. Researchers and industry experts need to come together in Australia and work closely to build stronger and more resilient capabilities that help safeguard businesses and users in today’s world of cyber war and cyber terrorism.

"We need to regain the community's confidence in our cyber capabilities.”

www.unsw.edu.au

ends