Artificial intelligence enjoyed a breakout year in 2023 as improvements to machine learning and natural language processing made AI much more practical to use, driving widespread adoption of generative AI chatbots like ChatGPT. In early November 2023, slightly less than a year after its launch, ChatGPT had amassed 100 million weekly users.
However, ChatGPT wasn’t the only game changer in the AI space in 2023. By making its LLaMA large language model available to the public free of charge, Meta also inspired an AI development boom. A leaked memo alleged to have been penned by a Google (News - Alert) researcher said the move by Meta impacted the AI community in a way “that cannot be overstated.”
“Both LLaMa and ChatGPT had significant transformational impacts this year, each in its own way,” explains Yashin Manraj, CEO of Pvotal Technologies. “ChatGPT generated a radical opinion shift on the importance of AI in our everyday lives and in business, especially during a heavy tech layoff season focused on automation and reduced workforce.”
Manraj has served as a computational chemist in academia, an engineer working on novel challenges at the nanoscale, and a thought leader building more secure systems at the world’s best engineering firms. His deep technical knowledge of product development, design, business insights, and coding provides a unique nexus to identify and solve gaps in the product pipeline. He currently leads the mission at Pvotal to build sophisticated enterprises with no limits that empower rapid change, seamless communication, top-notch security, and scalability to infinity.
“ChatGPT provided effective generative AI that could support straightforward business uses, including improving customer service and generating marketing copy, product descriptions, legal documents, music, poetry, and video games,” Manraj adds. “Certain activities such as information extraction and sentiment analysis allowed ChatGPT to optimize its impact by doing away with many of the unfortunate hallucination problems that appeared in competitors’ models in responses to complex questions.”
LLaMa inspired a new conversation among enterprise leaders
Compared to ChatGPT, LLaMa’s impact was felt more in the technical arena. LLaMa 2 was released in July 2023 as open source and free for commercial use, providing countless entrepreneurs with a foundation upon which to build AI-powered tools. It also caused a dramatic shift in the conversations surrounding AI development in the business world.
“LLaMa forced companies that were not already developing their own prompt or weight tuning to change their strategy rapidly and consider the limitations OpenAI had,” says Manraj. “In the corporate space, it prompted a significant shift among CIOs and CTOs to source away from the core OpenAI model and toward newer platforms and more customized implementations. Those leaders were facing the difficult task of weighing the risk of having their private data leak to a third-party model without guardrails, recourse, or guarantees. It became a race of innovation versus data integrity that has yet to find a winner between all available options. LLaMa opened a more technical and deeper dialogue for decision makers and business leaders.”
OpenAI inspired a major step forward in public policy targeting AI
In addition to its impact on the advancement of generative AI, ChatGPT developer OpenAI also helped to advance US government efforts to address the potential dangers of AI. An Executive Order issued in October 2023, which followed meetings between the White House and OpenAI, seeks to establish and address AI’s potential impact on the labor market, consumer privacy, civil rights, and other key issues.
“In theory, the Executive Order lays out guidelines meant to inspire large players in the AI space to cooperate and provide the US government with the results of AI safety tests and other critical information on development,” Manraj says. “It lays out an expectation that they will develop new standards, tools, and tests to help ensure AI systems are safe, secure, and trustworthy.”
2023 inspired an ongoing debate on the merits of open-sourced models
While LLaMa’s open-source model may have encouraged entrepreneurs to move deeper into AI development, it did not put an end to concerns among enterprises over data privacy and security.
“A new trend has emerged that involves malicious developers seeking to poison enterprise AI in an effort to facilitate a new era of scam, fraud, and automated theft of digital assets or services,” warns Manraj. “This could impact open source models more easily than closed source ones, in the short term. Many large companies seem to have given up on any plans to build or manage on-premise large language model training pipelines. They are exploring a cloud digital transformation prior to selecting the dominant enterprise AI model in the next few years, which could heavily favor closed-sourced companies when it comes to AI development and integration.”
As Manraj explains, the recent Executive Order could change the landscape by introducing new safety and security standards involving machine learning, AI, and large language models. Nevertheless, he warns there will always be factors that threaten the effectiveness of regulatory controls.
“We hope the recent government involvement will help foster cooperation within the industry to solve the systemic problems around large language model data integrity and privacy to protect the interests of the public,” Manraj says. “However, large malicious actors in China, Russia, and elsewhere won't be significantly impacted by such controls. At this point, there is no tangible solution available to protect against the risks of using AI to engineer dangerous biological materials, to enable fraud, drive discrimination, or create war propaganda, deep fake porn, and false documents.”
The developments of 2023 clearly revealed the power AI has to reshape the workplace and boost business productivity, but it also revealed the dangers that could accompany irresponsible applications of the technology. Moving forward, developers and businesses face the challenge of establishing a balance between using AI to drive innovation and mitigating the risks it poses.
