Is an Autonomous AI future possible?
Faster than Usain Bolt over 100 meters, AI has made lightning-quick strides over the last couple of years. Open AI and generative AI have hit the mainstream, with just about everyone jumping on board to produce their own bot or AI-driven tool designed to save time and money. In a utopian world, of course, humans would not need to do any energy-sapping manual labor. AI could do all the boring tasks for us with greater speed and accuracy, freeing humans up to do high-value work. Skeptics and doomsday prophets, on the other hand, are convinced that AI will leave millions unemployed and potentially empower malicious groups to fight wars with lethal and unlimited resources.?
If you have been following our series of articles on AI and Salesforce, you will probably have gathered that the truth lies somewhere in between the golden haze of AI optimists and those who fear the technology. Salesforce, while not being the first to the party, is certainly one of the most popular guests at the table. In a world where there are still many misgivings around data sources that underpin AI, and the security of AI systems, Salesforce shines as a beacon of reliability with its Einstein Trust Layer and commitment to ethical use. For many observers, it’s the magic combination of AI+CRM +Data+Trust that makes Salesforce so special and worth investing in during the AI revolution.?
Right now AI needs serious supervision - and many of us believe that certain high-priority tasks can only be conquered by humans. But what if this were to change in the near future? What would our Salesforce (a now AI-enhanced CRM), and our world, look like? Keep reading to find out.??
What is Autonomous AI?
But what is autonomous AI in any case? On a very high level, autonomous AI refers to the ability of AI to work with little to no human supervision. This work can include routine tasks but also complex decision-making, information processing, and learning based on experiences. Autonomous systems run by AI have the potential to change the world as we know it, and as of now some 77% of (human) workers apparently trust an autonomous AI future. This is according to a recent article by Salesforce dated June 26, 2024, which summarizes a double-anonymous survey they ran with YouGov earlier this year.?
This survey included almost 6000 workers from various companies and industries in the US, the UK, Ireland, Australia, Germany, France, India, Singapore, and Switzerland. And while a survey of this size is not necessarily comprehensive, it does offer interesting insight into how workers think AI is going to shape the future of work. Other findings include that these workers already rely on AI to do almost 43% of their tasks while leaders trust AI to do 51% of their work (versus regular employees to do 40%).??
AI Autonomy?
Diving a little deeper, workers believe the following about an autonomous AI future:
Some of the tasks respondents believe AI can be trusted with include writing code and communications as well as acting as a personal assistant and providing data insights. For now, it seems that tasks such as creating a sense of inclusivity at work, performing training and onboarding, and safeguarding data are perceived as mainly human skills.? Greater trust in autonomy in AI will naturally occur as workers upskill and gain more knowledge of how these processes work. It will also be more likely when we can all see examples of these tasks being executed successfully and efficiently without compromising security. The blooper reel of autonomous AI gone wrong is pretty lengthy. Who can forget the fatal crash involving a self-driving Uber in 2018, the embarrassing tweets by Microsoft’s bot Tay in 2016, and the bias found in Amazon’s AI recruitment a few years ago? Don’t forget AI can still be sexist and prejudiced in all kinds of ways.??
Salesforce AI
Salesforce as an organization is taking several steps to prepare for a world of autonomous AI in which AI agents become commonplace in the working environment. Some of these sophisticated agents may be built using Salesforce generative AI (Einstein Copilot?) technology and Salesforce Einstein AI more broadly, but first, people must trust it. Salesforce believes training and awareness around this technology is what will make the real difference. Keeping data models 101% accurate and trustworthy will mean workers will be far readier to fully rely upon artificial intelligence to make decisions on its own. There are still many barriers to making this a reality, one of the biggest being trapped and siloed data along with unreliable data, both of which issues Data Cloud and Einstein Trust Layer are trying to combat respectively.?
Furthermore, Salesforce is aiming to implement human checks and controls to autonomous AI systems to make sure they are working effectively and not going rogue. This includes ensuring that AI gets checked on at critical junctures, that it remains unbiased, that human beings interacting with AI are aware of this fact, and that AI doesn’t produce “Hallucinations” or false findings.?
Salesforce wants its workers to be prepared to work directly with AI agents to get the best possible results and ensure efficiency. Fast-forward 5 years and it’s going to be very interesting to see how roles like administrators change. Will the main skill set be determined by the admin’s ability to direct bots effectively? Everyone, welcome to management, we all got promoted!
Autonomy in AI: Wrapped
Thanks for reading! We hope you learned more about how autonomous AI is perceived in the workplace and how Salesforce is taking steps to ensure that it's ready for a world in which AI takes on work and makes decisions all by itself. It is clear that from Salesforce’s perspective, even autonomous AI should be heavily scrutinized and audited, but what happens when the bots start checking on us? Watch this space!