Microsoft sinks billions into OpenAI, but who is winning Big Tech’s AI arms race?

main__photo
As Microsoft pledges a new, multi-billion dollar investment in OpenAI, iGaming NEXT provides a rundown of some of the most exciting artificial intelligence applications from across the world’s largest tech companies.

Microsoft and OpenAI: Time to scale

Microsoft announced the third phase of its partnership with ChatGPT owner OpenAI yesterday (23 January), confirming it had made a “multi-billion dollar investment” in the firm which has this year taken the tech world by storm.

ChatGPT – first released in November 2022 – is a chatbot built on OpenAI’s proprietary GPT-3 family of large language models.

The product quickly separated itself from a herd of other AI-based chatbots due to its seemingly near-human command of language, allowing it to provide users with highly detailed and often surprisingly articulate answers to their requests.

While its core function is built around the mimicry of human conversation, the application brings a broad range of functionalities, including the ability to write and debug computer code, compose music, write poetry, format film scripts and even play games.

Microsoft’s latest investment in the firm follows on from previous investments in 2019 and 2021. The investment is likely to reach $10bn, according to Bloomberg.

In a blog post yesterday, Microsoft CEO and chairman Satya Nadella said: “In this next phase of our partnership, developers and organisations across industries will have access to the best AI infrastructure, models, and toolchain with [Microsoft cloud service] Azure to build and run their applications.”

Indeed, the latest investment round is intended to provide OpenAI with the ability to scale its products’ availability through the development and deployment of specialised supercomputing systems.

Occasionally at present, disappointed users of ChatGPT are being turned away and told that the product is operating at full capacity, as its popularity continues to grow.

Google: Taking it slow

Google parent company Alphabet, meanwhile, is taking a more cautious approach to the use of AI language models.

While the firm has recognised the undeniable buzz around products like ChatGPT, it has been careful not to rush into releasing a rival of its own.

Google CEO Sundar Pichai and head of AI Jeff Dean told staff at a meeting in December that this was due to the dangers of rushing a release of this kind of technology.

Google already has similar capabilities to what ChatGPT can deliver, the pair insisted, but equally, it has a more valuable reputation to protect and uphold than Silicon Valley-based start-up OpenAI.

As reported by CNBC, Dean said during the meeting it was “important to realise these models have certain types of issues,” namely that the accuracy provided by Google’s core search functionality cannot yet be reproduced inside a chatbot.

Indeed, even OpenAI CEO Sam Altman has recognised the limitations of a model like ChatGPT.

It is “incredibly limited,” he said on Twitter in December, “but good enough at some things to create a misleading impression of greatness”.

“It’s a mistake to be relying on it for anything important right now. It’s a preview of progress; we have lots of work to do on robustness and truthfulness,” Altman concluded.

It seems, therefore, that Altman shares at least some of Google’s concerns, namely that while the bot is able to produce impressively eloquent answers to queries, “the danger is that it is confident and wrong a significant fraction of the time.”

Despite the trepidation, do not assume that Google will let the product come to market unchallenged.

The search giant has been working on its own AI-based language model, LaMDA (among other AI applications) and is looking for further products to come over time, according to Dean.

CEO Pichai told staff in December that Google has a lot planned for AI in 2023, before concluding: “This is an area where we need to be bold and responsible, so we have to balance that.”

Meta’s CICERO: At the intersection

Facebook owner Meta, meanwhile, announced in November its new tool CICERO, “an AI agent that negotiates, persuades, and cooperates with people.”

This application exists at the intersection of two core areas of AI research: natural language processing, as used in models like ChatGPT, and strategic reasoning, used in products such as AlphaGo and Pluribus.

In 2019, Pluribus was declared the first AI bot capable of beating human experts in six-player no-limit Texas Hold’em, as it “decisively” took down professional poker players, including two World Series of Poker Main Event winners.

The combination of those two research areas led to the development of CICERO, which Meta announced in November was the first AI to achieve human-level performance in the strategy game Diplomacy.

This set the product apart from other AI models trained to beat humans in games with fixed variables such as chess, go, or any number of video games.

Playing an online version of Diplomacy, CICERO was able to achieve more than double the average score of human players and rank in the top 10% of participants who played more than one game.

Meta suggested this had long been considered a near-impossible goal for an artificial intelligence to surmount, “because it requires players to master the art of understanding other people’s motivations and perspectives; make complex plans and adjust strategies; and then use natural language to reach agreements with other people, convince them to form partnerships and alliances, and more.”

If an AI cannot recognise when players are bluffing, for example, or cannot predict how players are likely to perceive its own moves, it will fail in a subtle game such as Diplomacy.

Likewise, if it does not convince other players of its humanity, it will not be able to cooperate with them effectively enough to find success.

At present, CICERO is focused solely on playing Diplomacy – perhaps not the most useful application for the technology moving forward.

Underlying this ability, though, is the possibility to create several real world applications for the tech, Meta said.

It will likely be used to improve communication between humans and AI bots, for example by allowing for longer-form conversations during which an AI agent could teach a human a new skill – effectively becoming a near-human teacher for its users.

“We’re excited about the potential for future advances in these areas and seeing how others build on our research,” Meta said. 

Apple: The AI buyer

Apple was an early mover into the AI space but has been relatively quiet in recent years, especially compared to some of its competitors.

AI technology of course underlies the capabilities of its virtual assistant, Siri, which has been active since 2011, and also powers the firm’s FaceID technology, which allows users to securely unlock their devices.

Apple has undeniably adopted a “build and buy” approach to AI in recent years, acquiring 25 separate AI companies between 2016 and 2020 according to Brazilian tech journalist Filipe Espósito. Google, by comparison, acquired 14 over the same period.

The firm is taking a broad approach to AI functionality, acquiring a range of businesses such as home security camera start-up Lighthouse AI, autonomous vehicle firm Drive.ai, and AI Music.

While the firm continues to invest heavily in AI, major announcements have not been forthcoming in recent years.

Apple continues to advertise positions for employees working within AI and machine learning, in order to help build “amazing experiences into every Apple product, allowing millions to do what they never imagined.”

The business is currently advertising roles in machine learning infrastructure, deep learning and reinforcement learning, natural language processing and speech technologies, computer vision and applied research.

Earlier this month, the firm unveiled a suite of AI-voiced audiobooks, with a view to capitalising on the fast-growing medium, which is predicted to become a $35bn market by 2030.

Amazon: Integrating AI across a business

Amazon is undeniably an industry leader in the fields of artificial intelligence and machine learning. The firm uses the technologies to help solve a vast array of problems, from optimising processes in its warehouses to interacting with end-users via its Alexa virtual assistant.

The tech giant was an early mover in the space and now offers other businesses the chance to capitalise on its technology through Amazon Web Services (AWS).

Services include anomaly and fraud detection, as well as content personalisation and the reduction of customer churn for retailers, all of which use Amazon’s proprietary machine learning technology.

The company also relies on deep learning – a branch of machine learning that involves layering algorithms in order to better understand data – for use cases such as speech recognition and natural language understanding, image and video classification, and for powering recommendation engines.

Those uses all contribute to Amazon’s unparalleled success – with 56.7% of all US online retail purchases taking place via the retailer in 2021, according to Pymnts.

In October last year, the firm held an Innovation Day celebrating 20 years of experience in AI and machine learning (ML).

“The use of ML isn’t slowing down anytime soon, because ML helps Amazon exceed customer expectations for convenience, cost, and delivery speed,” the firm said in a related blog post.

While Jeff Bezos’s business has made AI and machine learning a core component of its success over the last two decades, using the technology to optimise its processes and reduce its costs, thereby delivering ‘death by a thousand cuts’ to its competitors.

The future’s not ours to see

The race to develop useful AI is most certainly heating up in 2023. While leading companies continue to help the technology scale, innovative start-ups like OpenAI are constantly pushing the boundaries.

The way we interact with technology is undergoing a paradigm shift, and it’s fair to say that AI will become an increasingly significant part of both our working and personal lives in the future. 

Earlier this month, iGaming NEXT investigated the gambling industry’s increased investment in AI, a developing trend as companies in the space prioritise profitability.

About the author

photo
Conor Mulheir

Conor entered the gaming industry in 2018 producing high-level live event content for audiences in London, Amsterdam and São Paulo. From 2020, he went on to report news and commission exclusive content for various gaming media brands before joining iGaming NEXT as editor in January 2022.

Subscribe to our weekly newsletters

Related Stories