Your data’s security is our top priority—and embedded into all our processes, which we work continually to keep improving.
While improving the quality of your analysis and keeping your client's data secure.
ArmsLength.AI utilizes OpenAI's models through secure Microsoft Azure API. While your inputs drive results, they are not used to train the AI.
Data is temporarily held on Azure servers, known for top-tier security. Once your session ends, this data is promptly cleared, ensuring no lingering information.
Our system encrypts all interactions via SSL. Additionally, ArmsLength.AI staff access data only upon specific user requests, maintaining a strong confidentiality standard.
The best money I have spent in my life. Thanks to this template, we saw a 120% increase in conversion rates and doubled revenue in three hours. I want to thank all the flowyak's team for making these results possible. Their only goal is to make small startups succeed and they are very good at it. We are the proof!
The best money I have spent in my life. Thanks to this template, we saw a 120% increase in conversion rates and doubled revenue in three hours. I want to thank all the flowyak's team for making these results possible. Their only goal is to make small startups succeed and they are very good at it. We are the proof!
Talk to our specialist about how we secure your data and ensure it's all risk-free.
How does ArmsLength AI arrive at its output?
The AI uses a combination of the base GPT model, the user's data, and the specific prompts/instructions from ArmsLength AI to generate its outputs. If, for instance, a user highlights an issue like the AI ignoring "credit risk" in an analysis, ArmsLength AI will adjust the instructions it provides to the AI to ensure it factors that in for subsequent tasks. The adjustment process involves humans reviewing the cases and fine-tuning the prompts rather than directly feeding the AI with raw user data.
How is the AI trained, especially concerning client-sensitive information?
The ArmsLength AI employs OpenAI's GPT base model through Microsoft Azure API. ArmsLength AI doesn't have its own AI but uses the GPT engine and directs it based on how the questions or prompts are posed to it. When a user interacts with ArmsLength AI, the base GPT model takes the provided user data and commands from ArmsLength AI to generate the desired output (e.g., a legal agreement summary). Importantly, this data isn't used to train the AI further, as it remains static.
There's a contractual understanding between ArmsLength AI and providers (Microsoft and OpenAI) which ensures user data is not used for training.
Does it use the information the user inputs that include client-sensitive information?
Yes, the AI uses information that users provide, such as the content of a legal agreement or context/parameters of functional analysis. This information, however, is not utilized for training purposes.
How is Client A's information kept confidential and not used to benefit Client B?
The information provided is temporarily stored on the ArmsLength AI server hosted on Microsoft Azure. It's not used automatically to enhance the solution. Only if a user indicates dissatisfaction with the results does ArmsLength AI review the case.
When making adjustments to improve results, they don't directly use raw client information. Instead, ArmsLength AI modifies the prompts and instructions they give to the AI to yield better outcomes in future tasks.
Can Client A’s information be anonymized?
Currently, ArmsLength AI doesn't anonymize the data because it is securely stored and not used for training by either ArmsLength AI or the underlying AI model providers (Microsoft or OpenAI).
Anonymization isn't deemed crucial, but it's open for discussion if clients feel the need for it.
How is the information secured?
Azure Security: Our system is built on the Microsoft Azure Web App, utilizing Azure's top-tier security infrastructure. This encompasses Azure DDoS Protection and advanced threat analytics, among other features.
Data Encryption: All data, whether in transit or at rest, is encrypted employing Azure's industry-standard encryption protocols.
OpenAI API Security: The Microsoft OpenAI APIs we deploy are well-fortified with a robust security suite. Importantly, this data is not repurposed for AI training. For an in-depth dive, refer to this page.
Where is the information held?
Azure Data Centers: We use Azure's internationally distributed, ISO-certified data centers which uphold stringent physical and digital security standards.
Information Storage: We don't retain user information for longer than 14 days. Concluding your session with our app ensures your input data will be purged from our servers. It is not used for AI training purposes and will only be accessed if the user is dissatisfied with the solution's performance. Exceptions are non-sensitive tools, such as Industry Analysis and Group Overview.
Redundancy and Backups: For maximum resilience and accessibility, our data undergoes regular backups across geographically distinct locations.
Who has access to the data?
By design of the process, ArmsLength.AI personnel don't access to user inputs or outputs and does not review them. Only when a user flags a performance anomaly, and they specifically request an investigation, do we delve into it.
The best money I have spent in my life. Thanks to this template, we saw a 120% increase in conversion rates and doubled revenue in three hours. I want to thank all the flowyak's team for making these results possible. Their only goal is to make small startups succeed and they are very good at it. We are the proof!
The best money I have spent in my life. Thanks to this template, we saw a 120% increase in conversion rates and doubled revenue in three hours. I want to thank all the flowyak's team for making these results possible. Their only goal is to make small startups succeed and they are very good at it. We are the proof!
Talk to our specialist about how we secure your data and ensure it's all risk-free.