{"id":12894,"date":"2019-06-12T09:00:04","date_gmt":"2019-06-12T09:00:04","guid":{"rendered":"https:\/\/www.microsoft.com\/en-gb\/industry\/blog\/?p=12894"},"modified":"2019-11-06T15:11:34","modified_gmt":"2019-11-06T14:11:34","slug":"ai-and-the-digital-ceo","status":"publish","type":"post","link":"https:\/\/www.microsoft.com\/en-gb\/industry\/blog\/cross-industry\/2019\/06\/12\/ai-and-the-digital-ceo\/","title":{"rendered":"AI and the digital CEO"},"content":{"rendered":"

\"Blogger<\/p>\n

A great CEO creates the environment for their team, their company, and their partners to succeed by making and influencing thousands of decisions. Some, like a decision to merge with a rival, will be huge and complicated with multiple decision makers. Others, like booking travel to make an important meeting on time, will be delegated. Many, like choosing the right corporate culture, will shape how others will make their own critical decisions.<\/p>\n

One of these decisions, often done in partnership with other decision makers such as a CTO, or CIO is when to bring technology into business. Whether it\u2019s from an operational or customer experience point of view, you should look at the outcomes you want to achieve before you look at the tools you need to complete this. In our deep dive into the state of the UK’s AI scene, ‘Accelerating the competitive advantage with AI<\/a>‘, we revealed that just 8% of businesses consider themselves as advanced AI users – yet it’s precisely this technology that can help CEOs make the decisions for their business.<\/p>\n

Why should a CEO consider AI?<\/h2>\n

\"AAI has great potential to bring the people in your organisation together. It extends your capabilities, frees up time to be more strategic and innovative and helps your organisation achieve more. The appeal of AI to decision makers is that it simplifies and reduces the number of decisions that we make. It even has the potential to provide better recommendations.<\/p>\n

On average, humans forget 80 percent of their teaching. When pressured to make decisions, humans regularly choose the approach that they have previously employed. In contrast, AI remembers all that it has previously encountered. It constantly updates its decisions based on the most recent data.<\/p>\n

We are already seeing AI use this ability to remember, recall, and revolutionise to succeed in computer and board games. In those cases, AI is rapidly learning the rules of the game and developing new strategies previously not considered by humans.<\/p>\n

At the same time, AI is learning to infer more information from incomplete information that it encounters. Microsoft\u2019s Semantic AI[1]<\/a><\/sup> demonstrates a fluid conversation between a machine and human, where the AI is determining the subject of conversation from relationships, partial names, and appointments.<\/p>\n

In both of these cases, the human receives the output (a perfect winning game strategy or a team meeting set up at the right time and place) yet do not immediately view the evidence or steps creating that output.<\/p>\n

These advances could tempt a CEO to replace their board with ever smarter AI. Why have a Chief Security Officer when AI can immediately address security risks? If AI has already predicted the optimum cost model and reorganised the finance structure during the discussion, why review financial performance? Is there a point to even having a discussion?<\/p>\n

The importance of ethics in AI<\/h2>\n

\"SmallMicrosoft is both excited and nervous by these advances. Excited, as we can empower more people and organisations to achieve more across the planet. Nervous, as we believe that AI should provide insight and transparency into how it has arrived at an output and that process should be shared with humans making the decision. We are not seeing this happen. When AI recommends a course of action for a critical decision, the decision maker must understand how that decision was made. A CEO needs to be able to scrutinise the recommendations from an AI Advisor in the same way that they will analyse and probe recommendations from a human advisor.<\/p>\n

This suggests a case to run competing AI advisors simultaneously. We are seeing humans increasingly reliant on the artificial recommendation, for instance following a satnav on a daily commute. We immediately see the catastrophic results when this goes wrong with faulty autopilots on aircraft. For business decisions, the consequences are usually slower to appear and, therefore, harder to attribute to a single poor AI recommendation. We need to be better advised, not better delegated.<\/p>\n

By using multiple competing AI recommendations, we reduce the amount of bias from a single system trained in a specific way, with specific data. Variety of thought is introduced back into the decision-making process. We reduce the risk of a catastrophic error from a single system. We stop lazy delegation to machines.<\/p>\n

\n

\u00a0Most crucially, we put the human at the decision point.<\/p>\n<\/blockquote>\n

Deciding between great recommendations and assessing relative benefits is where humans excel and is the essence of a successful team. AI offers an option for a CEO to remove the board and rely upon just a single AI advisor. Instead, we should be employing AI in the board room to improve the options for consideration rather than replace the decision makers.<\/p>\n