State of AI report predicts drop in funding for startups
One of the key findings from the State of AI Report 2022 was a predicted dry-up of the ‘megaround’ funding drives that we’ve seen for AI startups in previous years.
The report anticipates that the number of funding rounds that exceed $100mn will decrease significantly.
As it stands, the predictions are that AI startups will raise $70.9bn this year, which is a marked drop from the $111.4bn figure reached in 2021.
Key predictions for AI research, industry and politics
The State of AI Report, produced by AI investors Nathan Benaich and Ian Hogarth, has now released its fifth report. It explores the current status in a number of key areas, including research, industry, politics, safety and predictions.
Another interesting finding from the report was that new independent research labs are rapidly open sourcing the closed source output of major labs.
This is countering the long-standing dogma that AI research would become increasingly centralised among a small pool of players. Thanks to the lowered cost of and access to compute, pioneering research is coming from much smaller, previously unknown labs. But despite this, AI hardware remains strongly consolidated to NVIDIA (despite extensive investments coming in from Google, Amazon and Microsoft).
The report also revealed that the China-US AI research gap is continuing to widen. In fact, Chinese institutions are currently outpacing America’s at an exceptional pace, with 4.5 times as many papers being produced than America, and significantly more than the US, India, UK, and Germany combined.
Advancements in AI workplace safety standards
Safety in AI was another widely discussed topic in this year’s report.
It outlined how safety is gaining awareness among major AI research entities, as currently there are an estimated 300 safety researchers working at large AI labs.
This represents a marked growth, compared to the less than 100 working at the time of last year's report.
Alongside this, the rising recognition of major AI safety academics is a promising sign, indicating that AI safety is increasingly becoming a mainstream discipline.