OpenAI became the nexus of the technology world in 2023

We’re just over a year since it burst onto the scene and OpenAI’s ChatGPT program is somehow even more everywhere than it was in February. Our capability to regulate generative AI and mitigate its myriad real-world harms, on the other hand, continues to lag far behind the technology’s state-of-the-art. That makes 2024 a potentially pivotal year for generative AI in particular and machine learning in genera. ill AI continue to prove itself a fundamental revolution in human-computer communication, on par with the introduction of the mouse in 1963?, Or are we instead heading down yet another overhyped technological dead-end like 3D televisions? Let’s take a look at how OpenAI and its chatbot have impacted consumer electronics in 2023 and where they might lead the industry in the new year.

OpenAI had a great year, all things considered

“Meteoric” doesn’t do justice to OpenAI’s rise this year. The company released ChatGPT on November 30, 2022. Within five days, the program had passed 1 million users; by January, 100 million people a month were logging on to use it. It took Facebook four and a half years to reach those sorts of engagement numbers. ChatGPT outpaced the launches of both TikTok and Instagram to become the most quickly adopted program in the history of the internet in 2023. Heading into 2024, OpenAI (with billions in financial backing from Microsoft) stands at the forefront of the generative AI industry — whether the company can stay there, while billions more are being poured into its rivals’ R&D coffers, remains to be seen.

The company’s sudden success this year also launched its CEO Sam Altman into the media spotlight, with the 38-year-old former head of Y-Combinator basking in much of the praise formerly heaped upon Elon Musk. For a while, Altman was everywhere, repeatedly making appearances before Congressional committees and attending the Senate’s AI Safety Summits. He also conducted a 16-city world tour to Israel, India, Japan, Nigeria, South Korea, across Europe and to the UAE to help promote ChatGPT to developers and policy makers.

Even his termination at the hands of OpenAI’s board of directors in November ended up being a net positive. Fired on a Friday, Altman’s ouster set off 72 hours of panic in Silicon Valley with multiple OpenAI leaders resigning in solidarity, some 95 percent of rank and file staff threatening to walk without his reinstatement, the installation and removal of two interim CEOs in as many days and, ultimately, an indirect intervention by Microsoft. In the end, Altman is still CEO of OpenAI, now with a more compliant and agreeable board, and the tacit understanding throughout the industry that if you strike him down, Sam Altman will become more powerful than you can possibly imagine.

Keeping pace proved a challenge for OpenAI’s competition

A significant contributor to ChatGPT’s immediate and overwhelming success is that it was the first AI of its kind to market. Image generators like DALL-E and Midjourney were already popular diversions, and the public had long acclimated to more mundane machine learning tasks like language translation, but OpenAI was the first with a generative AI program that conversed naturally with its user. That novelty proved an invaluable advantage as even tech titans like Google and Amazon with their massive R&D budgets were caught unprepared for such demand and were slow to respond with competing products of their own.

Google was the most ignoble example of such imitators this year. Following ChatGPT’s debut, Google dedicated the vast majority of its I/O Developers Conference in March to a raft of brand new generative AI models and platforms, including the debut of the Google Bard chatbot. Bard was Google’s answer to ChatGPT, just not a particularly reliable one to start. Even before its public release, Bard made an embarrassing first impression when in February it confidently recited incorrect information about the James Webb Space Telescope in a Twitter ad.

Throughout the year, Google steadily added features, capabilities and access to Bard, eventually shunting the entire platform in December to its newly released foundational model, Gemini, which had been billed as Google’s “most capable and general model” built to date. Google was, of course, then immediately caught misrepresenting the system’s capabilities during a video demonstration. Even without once again getting caught in an easily disprovable lie, Gemini’s demo did little to quiet critics of Google’s stilted and frantic response to ChatGPT.

As a recent Bloomberg op-ed points out, yes, Gemini beat out ChatGPT in a majority of the industry’s standard performance benchmarks. However, Google used the as-yet unreleased Gemini Ultra model to earn its scores and the model only bested GPT-4 so by exceedingly narrow margins. GPT-4 came out nearly a year ago and Google’s best effort barely topped it in middle school-level algebra tasks. That’s not a great look from a corporation that boasts research budgets which rival the GDP of small nations.

Bing is doing just fine, thanks for asking. Microsoft dropped $10 billion on OpenAI in January as part of an ongoing multi-year partnership so now Bing — and literally everything else in the MS ecosystem — is being augmented with algorithmic intelligence. If there was one company that had a better 2023 than OpenAI, it’s Microsoft, which is reportedly set to receive 75 percent of all OpenAI’s profit until those invested billions are recouped.

Amazon placed its $4 billion generative AI bet on Anthropic’s Claude LLM, and made significant headway in leveraging the technology for use in its sprawling empire in 2023, from its Echo Frames smart glasses to Alexa with Generative AI to NFL Thursday Night Games. The company introduced its Bedrock foundational model platform (which will offer AI-generated text and images as a cloud service), launched a series of free AI Ready developer courses and an accelerator program to fund genAI startups, debuted generative tools for filling backgrounds and product listings and now offers a standalone image generator AI to rival DALL-E.

"Inside Amazon, every one of our teams is working on building generative AI applications that reinvent and enhance their customers' experience," CEO Andy Jassy said during the company’s Q2 earnings call in August. "But while we will build a number of these applications ourselves, most will be built by other companies, and we're optimistic that the largest number of these will be built on [Amazon Web Services]. Remember, the core of AI is data. People want to bring generative AI models to the data, not the other way around."

We’re still not ready for the age of AI

Even when it's not being used for obviously nefarious purposes like defrauding the elderly and amplifying political misinformation, generative AI technology has proven immensely disruptive to numerous industries and institutions from logistics and manufacturing to education and healthcare. It has been touted as a replacement for humans in professions ranging from medical imaging, computer programming and accounting to journalism and digital visual arts — in many cases, layoffs have been quick to follow.

This year also saw labor strikes by the Writers Guild of America and the Screen Actors Guild, in part, to prevent their works and likenesses from being used to train future AI models. Independent artists, whose intellectual property has been shamelessly scraped by disreputable firms for model training (looking at you, Stability AI), have had far less success in protecting their works — leading some creators to take drastic and damaging countermeasures.

Data privacy has proven a sticking point for AI companies in 2023. A ChatGPT bug found in March had apparently been sharing chat history titles (and potentially payment data). A trio of Samsung employees inadvertently divulged company secrets when they used ChatGPT to summarize the events of a business meeting in April. Microsoft AI researchers accidentally uploaded 38TB of company data to an open access Azure web folder in September, right around the time it was discovered that Google had been unknowingly leaking users’ Bard conversations into its general search results. As recently as November security researchers were finding that even “silly” attacks like telling ChatGPT to repeat the word “poem” ad infinitum would trick the system into revealing personally identifiable information.

The institutional response to these growing issues was tepid to start the year, mostly school districts, government agencies and Fortune 500 companies restricting use of chatbot AIs by their employees (and students). These initial efforts proved largely ineffective, due to the difficulty in actually enforcing them. The federal government's regulatory efforts are expected to have far more teeth.

The Biden White House has made AI regulation a centerpiece of its administration, developing a “blueprint” for its AI Bill of Rights last October, investing millions into new AI R&D centers for the National Science Foundation, wringing development guardrail concessions from leading AI companies and launching an AI Cyber Challenge, among other efforts. The administration’s most ambitious action came in October when the President issued a sweeping executive order establishing broad protections and best practices regarding user privacy, government transparency and public safety in future AI development by federal contractors. The US Senate and House have both been busy as well this year, holding congressional hearings on federal oversight rules for the AI industry, hosting a pair of AI Safety Summits and drafting legislation (which has yet to receive a vote).

Looking ahead to OpenAI’s 2024 and beyond

It’s OpenAI’s lead to lose heading into the new year. CEO Sam Altman holds firmer control over the company than ever, all dissenting voices on the board calling for caution have been silenced and the company is poised to further expand its operations in 2024 as the technology continues its global advance. I expect to see OpenAI’s competitors make a better showing in the new year with Google, Meta and Amazon spending freely on AI research in order to catch up and surpass the GPT platform.

And even though the entire ChatGPT craze got started with individual users, Paul Silverglate, vice chair of Deloitte LLP, sees the largest gains in 2024 coming from enterprise applications. “Expect to see generative AI integrated into enterprise software, giving more knowledge workers the tools they need to work with greater efficiency and make better decisions,” he wrote in a recent release.

A recent study by McKinsey & Company estimates that the current generation of conversational AI systems “have the potential to automate work activities that absorb 60 to 70 percent of employees’ time” thanks to rapid advancements in natural language processing technology with “half of today’s work activities" potentially being automated away from human hands "between 2030 and 2060." That’s a decade sooner than previously estimated.