For years, the overwhelming concern expressed by MTs regarding speech recognition was their jobs were at risk and that machines would “take over”. With the realization that speech recognition is technology to enhance MT performance and not a substitute for knowledge and experience, another challenge has risen – maintaining realistic expectations of draft quality and helping transcriptionists embrace their new role.
The following are common concerns/issues:
“I feel like I could type the job faster.” While the perception may be that editing is slower than transcribing, especially for newly-transitioned editors, it is extremely rare for this to actually be the case. It is important to keep in mind that you are moving from a fast-paced, physically-demanding environment to a slower-paced, mentally-demanding environment. This often plays a role in the perception of “slowness”. It is important to encourage efficient editing skills. Mentoring is also key to continued success. Do not be afraid to ask your trainers for additional help.
“I made the same edit several times and the machine is still not learning.” While seeing the same errors repeatedly can be frustrating, understanding how the speech engine learns may alleviate some frustration. It is important that you understand that the engine learns from statistically significant data. We would not want learning to occur in a short timeframe, as incorrect edits would negatively impact draft quality.
“What role do I play in the learning process?” There are numerous variables to successful learning, one of which is consistent editing by all MTEs. Therefore, the true role that an MTE plays in the learning process is to complete a top-quality document that can be distributed not only to your customer, but to M*Modal’s speech engine.
Do you have any concerns or questions that you would like to see addressed? Please leave us a message. We would love to hear from you!