Project Stargate is both fascinating and extremely frightening if you, like I did, grew up with movies like Colossus: A Forbin Project, WarGames, Terminator, and The Matrix. All of these movies foretold what would happen if we birthed a global-level AI badly. Stargate, named after a movie and series that had nothing to do with AI likely because Skynet would have been too much on the nose, is a $500B project with global implications.
As those movies showcased, a lot of thought and effort needs to go into assuring the ethics, quality and goals of the AI before it is turned on as it promises to become the biggest and most powerful Artificial General Intelligence (AGI) when it is completed, if OpenAI’s timelines are to be believed.
As a result, the company leading this effort must be one that knows large-scale monolithic deployments, understands AI from a realistic standpoint (in other words knows both its positive and negative potential and how to avoid the latter), and has demonstrated it can be trusted.
The Risks of Project Stargate
There are a large number of risks when bringing out a project as aggressive as Stargate appears to be. First, even if the project is completed flawlessly, which is unlikely given how new the technology will be and how complex it will become, it will be a huge magnet for bad behavior. The data it will have access to alone could be worth trillions to companies that want to exploit it, and hostile countries, should they gain control or are able to crash the AI, would have the power to bring the U.S. to its knees in a way that might be unrecoverable should Stargate be connected to much of the U.S.’s military with a growing number of autonomous systems, and given it will also be connected to U.S. infrastructure management.
Everything from utilities to traffic management could be disrupted were it hacked or poorly programmed, making the risks associated with this project potentially greater than its short-term benefits. This project would need to reflect all of what we’ve learned from AI, its tendency to become biased, its tendency to hallucinate data making for bad decisions, and its tendency to become corrupted by its users.
The level of security, quality control, checks and balances, and the critical need for an extremely secure off-switch should this thing go off the rails are all critical and need a company that understands the critical nature of these things to be in charge of the project.
Wrapping Up: Why IBM Should Be in Charge of Project Stargate
IBM is over a century old, and it is decades older than the companies currently planning to run Project Stargate. In addition, IBM started working on AI decades ago. Through its partnership with NVIDIA, IBM knows the technology likely to be used to create Stargate intimately.
Earlier this month I saw a presentation from IBM on its SAP partnership and the number of banks that continue to rely on IBM above all other providers, choosing Power and Mainframes over more common platforms because of a vastly greater need for security and portability of applications between on-premise and the cloud.
These are all things even more critical to Stargate, and IBM over and over again demonstrates that when security, quality, reliability and uptime are critical, and with Stargate, the word “critical” is inadequate to cover something that has global implications for the survival of the race, let alone U.S. competitiveness.
IBM appears to be the only company with the experience, reach, focus on ethics, focus on quality and with the culture to prioritize quality and safety over revenue and speed that is needed on this project.
IBM’s executives are formally trained, and IBM takes security far more seriously than their peers, making them the best, if not only, choice to run Project Stargate if we want it to improve the world, not end it.
