With a rash of artificial intelligence (AI) news making the rounds in the last month of 2023, opinions widely differ on how governments, companies and consumers will use the technology.
After months of deliberation, the New York Times officially filed suit against Microsoft and OpenAI saying their copyright was infringed to train ChatGPT.
Michael Cohen, Donald Trumps personal lawyer says he furnished his lawyers with bogus legal citations generated by Google’s Bard AI in sword declarations unsealed on Friday, December 29. In the unsealed statement, Cohen said he used Google’s Bard AI software for case citations, believing it was “a super-charged search engine,” and said he was not aware it was a generative AI (GAI) service like Chat-GPT.
And with Sam Altman back at Open AI, the company is in talks to raise funds at $100M or more.
Sustainability
Kaitlyn Albertoli, CEO and co-founder of Buzz Solutions cites sustainability as a market for impact.
“In 2024, we will see AI impact sustainability in two opposing ways,” said Albertoli. “On one hand, the massive energy consumption at the data center level required to power all of the emerging AI / GenAI applications and high power computing will have an ongoing negative effect on the environment, especially as the power grid is further strained.”
“And, on the other hand, AI will also be increasingly used to solve some of the most challenging issues with the environment, including algorithms that can empower organizations to reduce their carbon footprint by optimizing operations,” adds Albertoli. “In the energy industry, this includes things like AI for predictive maintenance and boosting the efficiency of renewable energy sources.”
Global challenges
Sian John, chief technology officer at NCC Group, says that AI provides a vital opportunity globally with great potential to enhance how we live and work.
“AI is being deployed across ways that impact all areas of our life, which brings significant risks if not controlled properly,” she said. “AI is a global challenge that needs a global response, and while AI offers endless opportunities, we must make sure that the adoption of AI practices is done responsibly and with cybersecurity in mind.”
“Major regulatory divergence between key nation states could be counterproductive and ultimately lead to poor security and safety outcomes for all,” adds John.
“The recent Bletchley Declaration – alongside the G7’s Hiroshima Process and domestic moves like the Whitehouse’s Executive Order – represent critical steps forward toward securing AI, on a truly global scale,” said John. “We are particularly heartened to see commitments from the Bletchley signatories to ensure that participants will convene again next year in South Korea and France for continued international leadership.”
John says by doing that, it will be important to set clear and measurable targets that leaders can measure progress against.
“There is a fundamental research challenge around energy use in the face of an energy crisis and climate change; it is estimated that 21% of the world’s overall energy usage by 2030 will be computing power alone, yet technologies in AI, cloud, and quantum are inherently power-hungry,” said John.
“Globally, we need to be incredibly mindful and proactive in developing secure, power-efficient and sustainable technical solutions for the systems we will create and use over the coming decade,” she adds.
Security
Joseph Thacker, principal AI engineer and security researcher at AppOmni says AI security will be increasingly focused in 2024.
“AI being used for security will also ramp up,” said Thacker. “There will be numerous startups, potentially hundreds, focused on AI security, and every major application will incorporate AI features, a shift that will inevitably introduce new vulnerabilities.”
“Startups will appear for many domains,” said Thacker. “There will be AI SOC analysts who will handle alert triage, AI ethical hackers tasked with uncovering vulnerabilities, AI code review tools capable of identifying software bugs and suggesting automated fixes, and AI social engineering toolkits designed to enhance phishing efforts, to name just a few.”
“Allowing AI systems to make decisions is convenient, and that means many products will incorporate it without adequate security testing,” said Thacker. “We will see where this leads really soon.”
Commerce
Chris Anley, chief scientist at NCC Group, takes a different view and says commerce will be the most important AI trend for 2024.
Anley says that 2023 saw a significant increase in the use of large language models (LLMs) like ChatGPT and Bard and Bing chat, which have now become household names.
“Organizations have struggled to keep up with the tricky balance of enabling staff to use these systems while keeping their sensitive information secure,” said Anley. “Commercial offerings like ChatGPT Enterprise, GitHub CoPilot Enterprise, and Amazon Q are specifically targeted at enabling enterprises to safely leverage this sensitive information via language models while retaining security controls.”
Anley says that in 2024, we can expect more businesses to leverage enterprise-oriented LLMs to improve productivity, drive efficiency, and provide new business capabilities and applications.
“Businesses are seeing that there are real financial benefits to be gained – perhaps not the science fiction style capabilities predicted by some, but certainly practical, measurable improvements in productivity,” said Anley. “Now that these benefits are becoming clearer – in financial terms – investment by companies in these systems will increase.”
Anley believes we will likely see a steady increase in using LLMs and AI assistants in business. “By the end of 2024, using AI assistants to help draft documentation, summarize documents, or reformat data will be as natural as using a spelling or grammar checker today,” he said.
“You can’t have commerce without trust, and you can’t have trust without security. The business benefits of AI come with risks,” said Anley. “Integrating sensitive information with LLMs will expose it to attackers in new ways; security controls that previously applied to (say) databases don’t apply in the same way to free-text training courses or Retrieval Augmented Generation systems. Making sensitive data more available and useful also exposes it.”
Anley says the increased use of AI as a security control also presents a challenge. “The efficacy of such solutions demands robust testing, evaluation, and assurance, lest we become too trusting of automated decision-making in security governance that we believe to be optimal but renders our systems more vulnerable,” he adds.
Enterprise
Varun Badhwar, CEO and co-founder of application security startup Endor Labs, predicts where AI will go from here.
“Right now, the world is at the intersection of AI and enterprise control,” said Badhwar. “Malicious actors will continue to explore using AI and associated tools to accelerate exploitation and intrusions, and they will also look to target the large GenAI platform providers and widely used AI OSS projects/components as part of broader software supply chain attacks.”
Badhwar says that in an evolving technological landscape, the parallels between the adoption of cloud services and the current surge in AI implementation are both striking and cautionary.
“Just as organizations eagerly embraced cloud solutions for their transformative potential in innovation, the haste of adoption outpaced the development of robust security controls and compliance tools,” said Badhwar. “Consequently, this created vulnerabilities that malicious actors were quick to exploit, leaving enterprises grappling with unforeseen challenges.”
“As we witness a similar trajectory in the adoption of AI technologies, it becomes imperative to draw lessons from the past and proactively address the looming concerns,” he said. “The rapid integration of AI into various facets of business operations is undeniably transformative, but the lack of comprehensive visibility and enterprise control raises red flags.”
Badhwar says businesses must prioritize the development of robust governance frameworks and visibility tools. “The focus should extend beyond the mere integration of AI solutions and encompass a holistic approach that addresses security, compliance, and ethical considerations,” he said.
“Learning from past experiences, organizations must strive to stay ahead of potential challenges, ensuring that innovation is not compromised by vulnerabilities,” said Badhwar.
“In an era where data is a prized asset, and AI is a powerful catalyst for business growth, the responsibility lies with industry leaders, policymakers, and technology providers to collaboratively build a foundation for responsible AI adoption,” adds Badhwar. “By establishing comprehensive controls and fostering transparency, we can unlock the full potential of AI while safeguarding the interests of organizations and their stakeholders.”
