WithSecure is currently the largest cybersecurity firm in the Nordics, promising outcome-based security for some of the largest financial institutions and manufacturers.
A 2023 GlobalData survey found that over 17% of businesses answered they had a high adoption rate of AI into their systems and AI is a topic garnering worldwide attention. Despite the recent buzz, Hintikka was quick to reaffirm that AI has been “around for a while” within universities and other research institutions.
“The advent of large language models (LLM) has kind of popularised the topic. I think there are plenty of people who find [LLMs] very useful in helping with their daily tasks. Using these AI models, like ChatGPT, for automating certain things, admittedly, is very useful,” Hintikka began.
On top of this leap of productivity, Hintikka also mentioned the growing ability of LLMs to understand and respond to languages outside of English.
“[AI] is learning more and more, even in Finnish,” he says, “personally, I’ve used it to write quick summaries and that it can do very well.”
However, this automation does, warns Hintikka, come with risk.
How well do you really know your competitors?
Access the most comprehensive Company Profiles on the market, powered by GlobalData. Save hours of research. Gain competitive edge.
Your download email will arrive shortly
Not ready to buy yet? Download a free sample
We are confident about the unique quality of our Company Profiles. However, we want you to make the most beneficial decision for your business, so we offer a free sample that you can download by submitting the below formBy GlobalData
“I think there are issues related to privacy, for example company sensitive information, that you need to consider everything you put in there,” Hintikka explained, “… I think these are issues that companies need to instruct their employees on.”
Whilst good digital hygiene can help businesses and employees protect personal data, Hintikka stated that it is common for organisations to be surprised at how much of their data is accessible online despite their good practices.
Hintikka is not alone in his privacy concerns.
Several large companies, such as Accenture and Deutsche Bank, have already banned the use of ChatGPT by employees citing anxieties over data privacy.
Hintikka on the balance of regulating AI
When considering the regulation of AI, Hintikka stressed that it is important to remember that AI is still in the “very beginning” of its development and that it is possible to fully understand the implications of AI already.
“The fact that we’re so much in the beginning makes it hard to regulate, because at the same time, I think we would like to ensure that we don’t over-regulate,” Hintikka continued, “… we shouldn’t try to stifle something which has the potential to become really big.”
Speaking specifically on European regulation procedures, Hintikka stated that there was often a “history of trying to regulate a topic once the game is already lost”, citing the regulation of Cloud computing as an example.
Additionally, Europe’s close neighbour, the UK, has public aspirations to become a global leader in AI regulation and is set to host an AI Safety Summit this November.
For Hintikka, the UK appears to be at an “interesting” crossroads in whether it follows a more European or American approach to regulating the technology.
“It’s a question of whether the UK find common ground with former partners in the EU, or whether it becomes more closely affiliated with the US,” Hintikka elucidated.
Adequate AI regulation, no matter the location, is in Hintikka’s eyes all about “balance”.
“We need to ensure that we have conditions in place that help provide transparency, accountability and audit trails, but at the same time we shouldn’t overdo it,” he stated.
Whilst Hintikka explains that WithSecure have not seen AI deployed on wide scale cybersecurity attacks so far, the CEO believed such attacks were possible in the near future.
“Previously, AI has been in the hands of the defenders, people like ourselves that use AI to recognise patterns and malicious content,” says Hintikka, but AI can also be appropriated by bad actors.
“AI has the potential of quickly providing an attacker with dozens of other variations of the same code, which makes it hard to detect,” he warns, “and we have seen some first examples of that. So, I think that is a concern that businesses need to be prepared for.”
Whilst an avid supporter of open-source software, Hintikka did see some opportunity for recent open-source AI software (such as Microsoft’s Llama) could be downloaded by bad actors with the intent to build large language model based attacks.
The inability to determine a downloader’s intent, was in Hintikka’s words, a “real concern” for regulators and developers alike.
Looking forward to the future of AI regulation, Hintikka believed that any legislation would undergo many adaptations in the future as we begin to see the further development of AI due to its unpredictability.
Collaboration and communication between regulators and tech companies is the key to successful regulation. WithSecure has already advised the Finnish government on multiple occasions to inform policymakers, and Hintikka is optimistic this collaboration will continue to be beneficial.
“At the moment,” he concluded, “it’s all about the dialogue because we are in the beginning of AI.”